You can measure all three in a single load test by collecting the right metrics at the same time:

1) Response time

Track how long each request takes.

Common metrics:

• Average response time
• p95 / p99 response time (more useful than average)
• Min/max response time

Example:

• “95% of requests completed in under 320 ms”

2) Throughput

Measure how much traffic the system handles over time.

Common metrics:

• Requests per second (RPS)
• Transactions per second (TPS)
• Bytes/sec if relevant

Example:

• “System sustained 850 RPS during the test”

3) Errors

Count failed requests and classify them.

Common metrics:

• Error rate (%)
• HTTP status code counts: 4xx, 5xx
• Timeouts
• Connection failures
• Business logic errors (failed validations, rejected transactions)

Example:

• “2.3% of requests failed, mostly 500 errors”

---

How to run it in one test

A. Use a load testing tool that captures all three

Popular options:

• JMeter
• k6
• Gatling
• Locust
• Artillery

These tools usually record:

• request duration
• request count over time
• success/failure status

B. Define the same request flow

Make sure the test simulates real user behavior:

• login
• search
• add to cart
• checkout
• API calls, etc.

C. Add metrics collection

At minimum, log:

• timestamp
• request duration
• response code
• success/failure

Then calculate:

• latency percentiles
• RPS/TPS
• error percentage

D. Add monitoring on the system side

Also collect server metrics to explain the results:

• CPU
• memory
• disk I/O
• network
• DB latency
• thread pool usage

---

Example test summary

A single load test report might look like:

• Test duration: 30 min
• Peak load: 1,000 users
• Throughput: 900 RPS
• Avg response time: 210 ms
• p95 response time: 480 ms
• Error rate: 1.8%
• Top errors: 500 responses, DB timeouts

---

Key tip

Don’t rely only on average response time.
To understand load-test quality, report:

• Throughput
• p95/p99 latency
• Error rate

That gives a much clearer picture of system performance.

If you want, I can also give you:

1. a k6 example script that measures all three, or
2. a JMeter setup guide.